Optimal System Design Overview

Principles

Definition

Optimal system design is a methodological approach concerned with identifying system configurations that maximize performance or minimize costs according to predefined criteria and constraints. This field investigates the complex interplay of system components, objectives, and trade-offs to achieve desired properties such as efficiency, robustness, or reliability, often employing analytical, computational, or optimization techniques.

Ontological type

Core Methods

Objective Functions

Uncertainty and Robustness

Key developments

Key Figures

Hierarchical Robust Optimization era

Michael Athans [1] emerges as a central figure in the Hierarchical Robust Optimization era (1973-1987), with affiliations tied to the Massachusetts Institute of Technology [2] and the University of California, Berkeley [3] during this period. In this era his key contributions include the LQG/LTR procedure for multivariable feedback control design [4], which integrated hierarchical planning concepts and robustness into control design, enabling design-centric explanations of system behavior. Athans [1] helped propagate hierarchical planning and robustness concepts into practical design frameworks, shaping later optimization-as-decision-support methodologies. This era's framing of optimization as a design-support tool is reflected in Athans [1]'s emphasis on explaining system behavior through models rather than merely computing optimal solutions [4].

Model-based Robust Synthesis era

In the mid-2000s the scenario approach to robust optimization, championed by Riccardo Calafiore and Massimo Campi, offered tractable finite-sample programs with probabilistic guarantees for uncertain systems. Avner Ben-Tal and Arkadi Nemirovski laid the robust optimization foundations, showing how worst-case uncertainty can be embedded in convex reformulations that support model-based design. Kennedy and O'Hagan advanced surrogate modeling and multi-fidelity design, introducing Bayesian emulation to propagate system-level objectives into subsystems while managing computational cost. Dimitri P. Bertsekas and Stephen Boyd helped translate these ideas into deployable engineering practice through scalable, hierarchical optimization frameworks and robust control concepts.

Convex Digital Resource Coordination era

Rajkumar Buyya [1] is a leading figure in the Convex Digital Resource Coordination era, affiliated with King Abdullah University of Science and Technology [2] and Hunan University [3] during this period. His 2011 paper, Optimal online deterministic algorithms and adaptive heuristics for energy and performance efficient dynamic consolidation of virtual machines in Cloud data centers [4], proposed online and adaptive strategies that reduce energy use while maintaining performance, illustrating the era's focus on scalable, real-time orchestration. Rajkumar Buyya [1] has been associated with King Abdullah University of Science and Technology [2] and Hunan University [3], reflecting the era's cross-institutional development of convex resource coordination. The energy-efficient dynamic consolidation focus highlighted in this work [4] laid the groundwork for scalable, provably efficient cloud-resource management that integrates convex optimization into live data-center operations.